You are given $N$ strings $S_1, \dots, S_N$, each consisting of lowercase English letters and having length $M$.
Initially, let $X = S_1$, and perform the following operation $N - 1$ times.
In the $i$-th operation, let $Y$ be the string formed by concatenating $X$ and $S_{i+1}$ in this order. Then, choose any contiguous substring of $Y$ of length $M$, and replace $X$ with that substring.
Output the lexicographically smallest string that can possibly be the final value of $X$.
Input
The first line contains integers $N, M$ in this order. ($2 \le N \le 2000, 1 \le M \le 2000$)
Each of the following $N$ lines contains a string $S_i$ of length $M$, consisting of lowercase English letters.
Output
Print the answer.
Examples
Input 1
2 3 cat cut
Output 1
atc
Input 2
2 1 a b
Output 2
a
Input 3
3 8 jastaway tatesoto soryuusi
Output 3
asoryuus
Note
In the first example, initially $X = \texttt{cat}$. In the first operation, $Y = \texttt{catcut}$. If we choose the contiguous substring from the $2$-nd to the $4$-th character, then $X = \texttt{atc}$, which is the lexicographically smallest possible.